Eye Image Processing for Iris and Pupil Occlusion Estimation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for eye image analysis, such as gaze estimation or eye tracking, face challenges in accurately determining the positions of the entire region of the iris and the entire region of the pupil due to eyelid occlusion, leading to inaccurate results, especially when the eye is in a partially closed state.

Innovation Solution

An image processing method that includes acquiring an eye image and its feature extraction result, determining predicted positions of the visible regions of the eyeball, iris, and pupil, and using ellipse parameters to accurately position the entire regions of the iris and pupil, while considering constraints like the independence of the iris and pupil from the eyelid and the occlusion effects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing methods are used to analyze eye images, then the processing can be performed, but the accuracy of determining the positions of the entire region of the iris and the entire region of the pupil deteriorates due to eyelid occlusion

Engineering Contradiction:

Improveaccuracy of determining positionsVSAvoideyelid occlusion

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent segments the eye image analysis into multiple independent prediction modules: one for the visible region of the eyeball, another for the entire region of the iris, and another for the entire region of the pupil. Each module processes specific constraints independently, allowing the system to handle occlusion by focusing on visible features while inferring occluded regions through geometric relationships and constraints.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent changes the parameters used for position determination by introducing ellipse parameters (center coordinates, semi-major axis, semi-minor axis, rotation angle) to represent the entire regions of the iris and pupil. This parameter transformation allows the system to infer the complete region positions even when partially occluded, by using the visible portion's geometric characteristics and the known spatial relationship between these regions.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If the eye is in a partially closed state, then the visible region is reduced, but the accuracy of eye movement analysis deteriorates

Engineering Contradiction:

Improvehandling partially closed eye statesVSAvoidaccuracy of eye movement analysis

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent employs dynamic constraint relationships that adapt to different eye states. The conditional joint module dynamically adjusts the prediction based on the visible region's characteristics and the inferred entire regions, allowing the system to maintain accuracy whether the eye is fully open or partially closed. The constraints are updated based on the actual visible features detected in the image.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent introduces intermediate prediction results as mediators between the visible region detection and the final eye movement analysis. The predicted positions of the entire region of the iris and the entire region of the pupil serve as intermediaries that bridge the gap between the limited visible information and the complete anatomical structure, enabling accurate eye movement calculation even when the eye is partially closed.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Device complexity

If constraints of eye anatomy are ignored, then the processing is simpler, but the accuracy of the results deteriorates

Engineering Contradiction:

Improveprocessing complexityVSAvoidaccuracy of position determination

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent implements a multi-functional target model that simultaneously satisfies multiple eye anatomy constraints: the independence of the iris and pupil from the eyelid, the spatial relationship between these regions, and the occlusion effects. By encoding these constraints within the prediction modules and conditional joint module, the system achieves high accuracy without requiring separate processing steps for each constraint, as they are integrated into the unified model.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20260030778A1Image processing method, device, and medium

Publication Date: 2026.01.29 BEIJING ZITIAO NETWORK TECH CO LTD

AI summary

The present disclosure provides an image processing method and apparatus, a device, a medium and a product. The method includes: first acquiring an eye image and a feature extraction result of the eye image; then determining a predicted position of a visible region of an eyeball, a predicted position of an entire region of an iris and a predicted position of an entire region of a pupil based on the feature extraction result; and then determining, based on these predicted positions, a predicted position of a visible region of the iris and a predicted position of a visible region of the pupil.